Algorithmic Collective Action with Multiple Collectives

2025-08-26

AI Summary

This paper addresses the limitation of existing Algorithmic Collective Action (ACA) research, which focuses exclusively on single collectives. We propose the first **multi-collective ACA theoretical framework**, modeling how multiple decentralized user collectives—each aligned on a common objective but differing in scale, strategy, and intervenable feature categories—coordinately influence a shared classification model via **feature version modification and signal injection** (i.e., bias signal insertion into overlapping or non-overlapping target classes). Our approach integrates formal theoretical modeling with empirical analysis to quantify the coupled effects of collective scale, objective alignment, and signal interaction. Key contributions include: (1) the first formalization of multi-collective coordinated intervention mechanisms; (2) identification of a fundamental trade-off between collective scale and alignment degree; and (3) development of an interpretable model of cooperative effects, establishing a novel paradigm for decentralized, multi-objective, user-driven algorithmic governance.

Technology Category

• Multiagent Systems: Mechanism Design
• Humans and AI: Crowd Sourcing and Human Computation
• Game Theory and Economic Paradigms: Coordination and Collaboration

Application Category

• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Responsible Web: Algorithmic accountability and transparency on the web
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

As learning systems increasingly influence everyday decisions, user-side steering via Algorithmic Collective Action (ACA)-coordinated changes to shared data-offers a complement to regulator-side policy and firm-side model design. Although real-world actions have been traditionally decentralized and fragmented into multiple collectives despite sharing overarching objectives-with each collective differing in size, strategy, and actionable goals, most of the ACA literature focused on single collective settings. In this work, we present the first theoretical framework for ACA with multiple collectives acting on the same system. In particular, we focus on collective action in classification, studying how multiple collectives can plant signals, i.e., bias a classifier to learn an association between an altered version of the features and a chosen, possibly overlapping, set of target classes. We provide quantitative results about the role and the interplay of collectives' sizes and their alignment of goals. Our framework, by also complementing previous empirical results, opens a path for a holistic treatment of ACA with multiple collectives.

Problem

• Addressing multiple collectives in algorithmic collective action
• Studying coordinated data changes across different user groups
• Analyzing collective size and goal alignment effects

Innovation

• Theoretical framework for multiple collectives
• Plant signals to bias classifier learning
• Analyze collective sizes and goal alignment